Comparative gallium-68 labeling of TRAP-, NOTA-, and DOTA-peptides: practical consequences for the future of gallium-68-PET

With the commercial availability of ⁶⁸Ge/⁶⁸Ga generators, cyclotron-independent on-site production of tracers for positron-emission tomography (PET) has become widely feasible [1, 2]. Thus, in the near future, a ubiquitous implementation of PET and PET/CT even in regions with less well-developed infrastructure can be expected, similar to the global story of success of ^99mTc-based scintigraphy which started half a century ago with the introduction of ⁹⁹Mo/^99mTc generators [1, 3]. In the long run, a partial substitution of single photon emission computed tomography (SPECT) by PET (and SPECT/CT by PET/CT, respectively) appears to be a realistic scenario in view of the advantages of PET, such as superior spatial resolution and sensitivity. Besides, in contrast to reactor-produced ⁹⁹Mo, ⁶⁸Ge is cyclotron-produced. This can be considered advantageous with regard to the recent insufficiency of global reactor capacity for reliable ⁹⁹Mo supply [4], and independence of ⁶⁸Ga-PET from nuclear reactors might positively influence the bias of its public perception.

Generally, ⁶⁸Ga labeling is done by complexation of the ⁶⁸Ga³⁺ ion. For this purpose, dedicated chelators usually have to be introduced into precursor molecules by bioconjugation, wherein they readily determine the labeling chemistry. To facilitate global implementation of ⁶⁸Ga-PET, production of ⁶⁸Ga radiopharmaceuticals must be simple, robust, and reliable; this demands highly efficient labeling chemistry and, therefore, highly efficient chelators. Recently, we have shown that the bifunctional triazacyclononane-phosphinate (TRAP) ligand [5‐8] possesses markedly improved ⁶⁸Ga labeling properties [6]. This applies also to TRAP-based peptide conjugates, the practical consequences of which we further elucidate in this study.

TRAP(RGD)₃ was prepared as described before [6]. NODAGA-cyclo(RGDyK) (‘NODAGA-RGD’) was purchased from ABX GmbH (Radeberg, Germany). DOTATOC was obtained from Bachem (Bubendorf, Switzerland). Fully automated ⁶⁸Ga labeling was performed using unpurified eluate fractions of a ⁶⁸Ge/⁶⁸Ga generator with SnO₂ matrix (iThemba LABS, Somerset West, South Africa), as described previously [6, 9] (5 min reaction at 95°C, pH adjusted with HEPES, pH 3.2 for DOTATOC and NODAGA-RGD, pH 2 for TRAP(RGD)₃, purification using C8 SPE cartridge). Radiochemical yield was calculated from decay-corrected product activity in relation to the sum of significant decay-corrected residual activities contained elsewhere, that is, in reaction vial, SPE cartridge, and non-product cartridge purging liquids.

Product activities (A _P) were measured after the end of preparation (approximately 15 min after the start of syntheses) and decay corrected to a typical injection time, 30 min after the start of synthesis (A _P,30). In order to be able to calculate corresponding specific activity values that are representative for the respective precursor amounts and independent from small deviations in the starting activity A ₀ (in our experiments, ranging from 800 to 1,050 MBq, depending on the regeneration state of the ⁶⁸Ga generator), product activities were normalized to a representative starting activity A _N = 1 GBq, according to A P , 30 , N = A P , 30 A N A 0. Specific activity (A _S) values were calculated by the division of A _P,30,N by the precursor amount used. It is assumed that all precursor peptide is actually retained on and subsequently eluted from the cartridge, and thus transferred into the formulation. This means that both retention and elution efficiency are considered 100%, both of which can be somewhat lower in practice. As a result, all given A _S represent the lower bounds and will never overestimate actual values.

Although previous comparative studies focusing on the basic chelator structures TRAP, 1,4,7-triazacyclononane-triacetic acid (NOTA), and 1,4,7,10-tetraazacyclododecane-tetraacetic acid (DOTA) already proved superior Ga³⁺ complexation/⁶⁸Ga labeling properties of TRAP [6, 7], these data are not sufficient to quantify the behavior of respective peptide conjugates, for two reasons: Firstly, functionalization of chelators with peptides, resulting in conjugates with a multiple of the molecular weight of the neat chelators, is definitely prone to change overall complexation properties with potentially unpredictable outcome. Secondly, the chelating moiety in compounds commonly dubbed ‘DOTA-peptides’ is actually not DOTA, but DOTA-monoamide (see Figure 1), which exhibits different Ga³⁺ complexation behavior [10]. In contrast, for TRAP and the bifunctional NOTA-derivative NODAGA, the structure of the chelating site is not affected by conjugation. To assess the impact of these effects, representative peptide conjugates (TRAP(RGD)₃ [6] and the commercially available ‘NOTA’- and ‘DOTA’-peptides NODAGA-RGD and DOTATOC, respectively, see Figure 1) were labeled under similar conditions using our standard automated procedure [6, 9].

Figure 2 shows that TRAP(RGD)₃ allows to use much lower precursor concentrations for labeling than required for NODAGA-RGD and particularly DOTATOC, which is why ⁶⁸Ga-TRAP(RGD)₃ can be prepared with much higher A _S (see also Table 1). Using 0.1 nmol of TRAP(RGD)₃, almost 5,000 GBq/μmol was reached with a satisfying decay-corrected yield of 66 ± 6%. The use of even lower amounts of TRAP(RGD)₃ (17 pmol) frequently resulted in preparations with extremely high A _S of >10,000 GBq/μmol, although not reliably reproducible. The highest A _S value observed during these experiments was 14,900 GBq/μmol (actual value, not normalized to starting activity), which is approximately 1/7 of the theoretically possible maximum value, that of carrier-free ⁶⁸Ga. Although such high specific activities are not usually needed for clinical applications, we nevertheless, deem this feature of high practical value for the following reasons:

Furthermore, regarding Figure 2, one notices that variation of radiochemical yields, reflected by the size of error bars, is the larger the lower precursor amounts are. This is because the generator eluate usually contains traces of ionic contaminants, such as Zn²⁺, Sn⁴⁺, Al³⁺, and Fe³⁺, the concentration of which in the individual eluates is varying. These can compete with ⁶⁸Ga³⁺ at the chelating site of the precursor, which is naturally the more impacting on labeling yield the lower the stoichiometric excess of precursor over ⁶⁸Ga³⁺ ion is. The error bars in Figure 2 show that except for precursor amounts exceeding 20 nmol, use of a TRAP peptide will result in a more reliable radiosynthesis, being less prone to be perturbed by variation of other parameters (reaction pH, eluate volume, trace metal contaminations, etc.). Differences in radiochemical yield and reproducibility are very pronounced for peptide amounts in the range of 1 to 10 nmol (e.g., equivalent to 1.4 to 14 μg of DOTATOC) which are frequently used in routine ⁶⁸Ga labeling procedures; near-quantitative yields and excellent reproducibility can be expected here for a TRAP peptide. This is of high relevance for routine GMP tracer production, where reproducibility and robustness of procedures is crucial. In addition, we assume that due to higher labeling efficiency, realization of kit labeling procedures will be much simpler using TRAP conjugates, which we deem of importance for the aforementioned possibility of global implementation of ⁶⁸Ga-PET. Finally, the recent introduction of NOPO, a TRAP variety designed specifically for monoconjugation, expands the portfolio of P-functionalized triazacyclononane-triphosphinate chelators, thus offering even more synthetic possibilities for development of ⁶⁸Ga tracers [15].